not planned Unreal Engine 5 viable for DCS?

[Bog9y]

Hi, 

 

I read in the FB group that the Unreal ENgine 5 is amazing and what DCS needs. I am clueless with software development , not an IT guy or engineer so was just curious if this is something that could be used by ED to make DCS even better? It sure looks amazing: 

 

 

[Aarnoman]

No. DCS does not use an older unreal engine, therefore it would take a complete remake to switch engines. This is not feasable or realistic.

[cfrag]

1 hour ago, Bog9y said:

I read in the FB group that the Unreal ENgine 5 is amazing and what DCS needs.

 

That is an intriguing statement. No doubt UE5 is an impressive achievement and amazing in many ways. The second part of that sentence is unclear to me:

it would be interesting to hear what parts of UE are filling a need for DCS. Game engines like Unreal, Unity etc are sandboxes for developers and they have individual strengths and weaknesses. They are by definition a jack-of-all-trades (since they need to cater to all game genres from FPS to Card Collecting). They provide a world to place your game in, strong programming support, some physics processing, audio, and usually strong graphics processing. Since they are a jack of all trades, they are also, like the saying goes, a master of none. 

 

DCS on the other hand specializes on strong physics modelling with support for gigantic terrain and ho-hum graphics and programming support. They are a master of this (flight sim) arena, but have lackluster performance in many others (graphic fidelity, audio processing, programming support). Switching to another game engine would theoretically be possible (at immense expenses), but the question is: what for?

 

The 'what DCS needs' part is unclear to me. What does DCS need today that UE specifically brings to the table? Better graphics? Graphic fidelity plays (to me) second fiddle to the flight model. DCS's network gaming support is exceptional, and it is doubtful that UE can add anything significant here. So, what would switching to UE gain? Since UE also has drawbacks, what would we need to give up when switching to UE - and would that be wise?

 

So, what parts of DCS need improvement, and why will UE fix that (and not, for example, CryEngine)?

Edited July 1, 2021 by cfrag

[Tank50us]

As good as unreal is, one first has to remember that the engine DCS currently runs on is one that was made by Eagle Dynamics, so switching to Unreal would effectively be a slap in the face to literally dozens of years worth of work. There's other issues as well, one of which was brought out by Ace Combat 7: Unreal is really good for shooters, and 'small environment' games, but not much else beyond that.

 

I think the best thing for ED to do when it comes to engine work is overhaul the current engine, and rebuild it for multi-core support. This will allow for the fine rendering we're used to, without some people running into the situation where their computers get hot enough to fry eggs on them. On top of that, if ED can get even more support from the community in the realm of quality of life improvements (much in the same way ArmA has benefited from its mod community), I think DCS will become a much better title as time goes on.

[Silver_Dragon]

4 hours ago, Bog9y said:

Hi, 

 

I read in the FB group that the Unreal ENgine 5 is amazing and what DCS needs. I am clueless with software development , not an IT guy or engineer so was just curious if this is something that could be used by ED to make DCS even better? It sure looks amazing: 

 

 

ED has own engine building from scrach (EDGE) and none talk about change to another 3D engine.

Edited July 1, 2021 by Silver_Dragon

[QuiGon]

4 hours ago, Bog9y said:

Hi, 

 

I read in the FB group that the Unreal ENgine 5 is amazing and what DCS needs. I am clueless with software development , not an IT guy or engineer so was just curious if this is something that could be used by ED to make DCS even better? It sure looks amazing:

 

Nope, DCS uses its own inhouse engine, which is much better suited for the scale of DCS than the Unreal Engine.

[Bog9y]

On 7/1/2021 at 11:36 AM, cfrag said:

 

That is an intriguing statement. No doubt UE5 is an impressive achievement and amazing in many ways. The second part of that sentence is unclear to me:

it would be interesting to hear what parts of UE are filling a need for DCS. Game engines like Unreal, Unity etc are sandboxes for developers and they have individual strengths and weaknesses. They are by definition a jack-of-all-trades (since they need to cater to all game genres from FPS to Card Collecting). They provide a world to place your game in, strong programming support, some physics processing, audio, and usually strong graphics processing. Since they are a jack of all trades, they are also, like the saying goes, a master of none. 

 

DCS on the other hand specializes on strong physics modelling with support for gigantic terrain and ho-hum graphics and programming support. They are a master of this (flight sim) arena, but have lackluster performance in many others (graphic fidelity, audio processing, programming support). Switching to another game engine would theoretically be possible (at immense expenses), but the question is: what for?

 

The 'what DCS needs' part is unclear to me. What does DCS need today that UE specifically brings to the table? Better graphics? Graphic fidelity plays (to me) second fiddle to the flight model. DCS's network gaming support is exceptional, and it is doubtful that UE can add anything significant here. So, what would switching to UE gain? Since UE also has drawbacks, what would we need to give up when switching to UE - and would that be wise?

 

So, what parts of DCS need improvement, and why will UE fix that (and not, for example, CryEngine)?

 

Hi, thanks for that detailed reply. Like I said, I'm just relaying what I read on FB written by someone. That's why I posted the question to ask people who are probably way more clued up on this stuff than I am. The person never stated why DCS would need UE, I suspect the statement was just based on the footage he may have seen of UE on youtube, it does look pretty impressive. 

Cool, thanks for the replies guys. It will be amazing to see where DCS will be in a year or 2 from now. Hopefully VR performance and image quality will improve. 

[J-1775]

On 7/1/2021 at 12:49 PM, Tank50us said:

I think the best thing for ED to do when it comes to engine work is overhaul the current engine, and rebuild it for multi-core support.

One sentence ─ all said! I hope DCS devs closely watch what Asobo did with upcoming Update 5 of MSFS2020. Although the PC version remains (currently) on DX11 they obviously managed to get a HUGE fps improvement by just sorting and 'ironing' out their code. One example mentioned is a dedicated core for the glass cockpit.
Finally, watch my mouth, DLSS will be the future of all games and simulations. In a few years there will be no VR simulation without that option. DCS, better prepare for that. Even if it means DX12...

[SharpeXB]

9 hours ago, J-1775 said:

DLSS will be the future of all games and simulations

Dynamic Learning UpScaling = DLUS 

Edited July 3, 2021 by SharpeXB

[Mars Exulte]

10 hours ago, J-1775 said:

Finally, watch my mouth, DLSS will be the future of all games and simulations.

  DLSS is adaptive anti-aliasing. That's literally it. It does not yield any magic gains, full stop.

 

10 hours ago, J-1775 said:

In a few years there will be no VR simulation without that option.

  I'm 100% certain there will be.

 

10 hours ago, J-1775 said:

DCS, better prepare for that. Even if it means DX12...

  It doesn't. AMD already released their own opensource version of the same thing, cause.

 

@OP

  Beware flashy sales ads. They're usually not relevant to real world application.

 

  As for game engines being interchangeable, in short, they're not. Just like in vehicles, while physically possible to change a motor, some motors are suitable to one task but not another. An engine suitable for a passenger car is useless for a heavy truck and vice versa. In the case of a flight sim, the engine needs to be optimised to handle the scale, speed, and physics of flight, which not all engines are designed for or appropriate for.

 

 Can any engine be rebuilt to purpose? Yes, of course, with enough labor and redesign : see Star Citizen which took a session based shooter engine and rerigged it to use as a open world MMO flight engine. Is this usually a good idea? No, also see Star Citizen which spent large sums of money and uncountable man hours hammering a square peg into a round hole instead of designing or using something more suitable in the first place.

[SkateZilla]

The DCS Graphics Engine is being moved to Vulkan, which would remove all the DX11 API Issues.
Vulkan API gives ability to use things that everyone asks for, that isnt possible on DX11 (ie RayTracing, FSR, Etc etc)

As Impressive as that Terrain Physics Video is, Look at the Draw Distances of less than 50 Feet for most of the objects.
It's not Feasable for a simulation like DCS.
 

Edited July 3, 2021 by SkateZilla

[Tank50us]

26 minutes ago, SkateZilla said:

The DCS Graphics Engine is being moved to Vulkan, which would remove all the DX11 API Issues.
Vulkan API gives ability to use things that everyone asks for, that isnt possible on DX11 (ie RayTracing, FSR, Etc etc)

As Impressive as that Terrain Physics Video is, Look at the Draw Distances of less than 50 Feet for most of the objects.
It's not Feasable for a simulation like DCS.
 

 

 

Yup. It's fine for something like a shooter, which is what Unreal is best for, but for a hard sim like DCS, it's not that viable.

[J-1775]

On 7/3/2021 at 6:28 AM, SkateZilla said:

The DCS Graphics Engine is being moved to Vulkan

Hooray! This really is good news ─ for VR users probably one of the best in years! And among the many stuff possible with Vulkan is DLSS.

[J-1775]

On 7/3/2021 at 6:01 AM, Mars Exulte said:

Beware flashy sales ads. They're usually not relevant to real world application

When I talk about DLSS I am not speaking of sales ads. I speak about my own personal EXPERIENCE in War Thunder (which I keep installed for several reasons, progress of the Dagor engine is one them). For VR users the introduction of DLSS was a void promise for a long time (two big updates). So I had no hopes at all, when I ticked it on their last update. But woosh, there it was: not alone the promised and expected improvement in FPS, the bigger suprise to me was the picture quality: a rock-steady sharp picture with almost all 'shimmering' and flickering gone. It was like magic. I don't understand why so many people misjudge DLSS as just another method of AA. Sure, AA is part of DLSS, but AA is only one column of its architecture.

[cfrag]

1 hour ago, J-1775 said:

 For VR users the introduction of DLSS was a void promise for a long time (two big updates). So I had no hopes at all, when I ticked it on their last update. But woosh, there it was: not alone the promised and expected improvement in FPS, the bigger suprise to me was the picture quality: a rock-steady sharp picture with almost all 'shimmering' and flickering gone. It was like magic. I don't understand why so many people misjudge DLSS as just another method of AA. Sure, AA is part of DLSS, but AA is only one column of its architecture.

 

That is really, really encouraging to hear, as my VR is maxing out my GPU, and at the current price points I feel no urge to order a new one from a scalper. 

 

I'm surprised as you to hear that DLSS also works as a good form of AA. I can only speculate that the scaling that DLSS is based upon works as a form of temporal AA. In any event, now I'm really looking forward to the time if/when DLSS or it's siblings come to DCS. Until then, I've temporarily renamed Syria to 'the shimmering flats' (WoW reference some of you may understand). 

[Nomatic]

9 hours ago, J-1775 said:

Hooray! This really is good news ─ for VR users probably one of the best in years! And among the many stuff possible with Vulkan is DLSS.

To put this bluntly….

 

I can NOT F**CKING wait!

 

This and multi-threading are (hopefully not hyperbole) going to be game changing. LOL 

[Apar]

On 7/1/2021 at 8:14 PM, QuiGon said:

 

Nope, DCS uses its own inhouse engine, which is much better suited for the scale of DCS than the Unreal Engine.

 

I agree UE5 can do more the the scale of DCS gameplay

[Xilon_x]

gravity is not realistic those rocks so heavy they move as if they were fake polystyrene rocks.

I would like an engine that can calculate whether when you build a paper plane it is flyable or non-flyable. gravity must be realistic.

Edited July 23, 2021 by Xilon_x

[Tippis]

27 minutes ago, Xilon_x said:

gravity is not realistic those rocks so heavy they move as if they were fake polystyrene rocks.

I would like an engine that can calculate whether when you build a paper plane it is flyable or non-flyable. gravity must be realistic.

That's not really  an engine issue, nor is the movement of those rocks really a matter of setting the right gravity constant — it's more a matter of how their mass is calculated and what forces are applied to them as terrain is manipulated around them and as they interact with each other. If anything, given how quickly they accelerate, the gravity sim looks like it's running at overspeed or with too high a gravity constant…

 

Not that it matters anyway, since these kinds of solutions are handy for exactly what is being shown off here: map modelling on a small scale. Flight sims pretty consistently don't run on generic engines because of the pretty specific demands they have — small-scale detail is never one of them.

Edited July 23, 2021 by Tippis

[Goffik]

The only way we have all the complexity and detail we have in DCS is because it uses it's own engine. Try to do that in a general game engine, especially one you didn't create yourself, and it just won't work. Right off the bat, for any added functionality for new systems you'll be waiting on the owner of the engine to do what you need done. Their priorities are not going to match with yours, especially for a general engine which is not specifically designed for flight simulations. And that's just problem one on an incredibly long list of other issues. Just because Unreal is capable of creating pretty graphics in a first-person-shooter does not mean it is any good for anything else, especially something with the complexity and scale of a flight simulation.

[Csgo GE oh yeah]

Maybe already mentioned, but i think i've read multiple times that the Unreal engine is not really suited  for huge terrains. 

[Xilon_x]

Sono affascinato da questi tipi di motori perché sono vere e proprie distruzioni che utilizzano fisica e dinamica. in un simulatore di volo si usano immagini 3D mentre qui in questo motore si usano i VOXEL, che sono piccoli sprite 3D e pesano ancora meno sul funzionamento del motore. immagina che realizzare un simulatore di volo usando piccoli sprite 3D (VOXEL) sarebbe un motore completamente diverso dagli altri motori comuni.

 

 

 

Edited July 23, 2021 by Xilon_x

[DaWu]

This thread lol. No you can’t switch engines like ur underwear. 

[IkarusC42B Pilot]

-_-

[Xilon_x]

8 hours ago, DaWu said:

Questo thread lol. No, non puoi cambiare motore come la tua biancheria intima. 

 

no I'm changing tread but I'm talking about the type of graphics engine and programming technique.

they are just demonstratives they will not be perfect but they are simple and intuitive.